Although the psychophysiological signs of fatigue are well known, automatic methods for the detec... more Although the psychophysiological signs of fatigue are well known, automatic methods for the detection of fatigue in employees in specific working conditions are still lacking. Many people do repetitive work on computers and become fatigued; therefore, the detection of fatigue in employees can help prevent accidents and increase their work efficiency. In this article, we propose an algorithm for the effective detection of fatigue which is based only on electrooculographic (EOG) signal. Three features were assessed: blink duration, blink amplitude, and time between blinks. To cause fatigue, the ${N}$ -back test, lasting for 60 minutes, was carried out. The article presents the research results for 24 users. The effectiveness of the proposed system was measured by the accuracy of classification. The average classification accuracy was 0.93 for user-dependent mode and 0.89 for user-independent mode. The results of the conducted experiments indicated that assessing the three proposed features can help in the effective detection of fatigue in users.
2019 IEEE 20th International Conference on Computational Problems of Electrical Engineering (CPEE)
The article presents a gender identification based on speech signal with supervised machine learn... more The article presents a gender identification based on speech signal with supervised machine learning implementation. At first, a database of speech signals in Polish language was collected. Next, a set of features from audio signal were calculated. The features were farther used to train a neural network. Audio signal processing and implementation of the neural network were performed in Python, and the calculation of features in the R language. Neural network training process was carried out using only CPU, then CPU with GPU and the times of the programs execution were compared. The obtained accuracy of gender recognition was 92.4%. The use of GPU accelerated the network learning process several times.
This article describes a vision system that uses deep learning to recognize 24 static signs of th... more This article describes a vision system that uses deep learning to recognize 24 static signs of the American Sign Alphabet in real time. As part of the project, images of signs from four publicly available databases were used as a training set. A DenseNet was implemented for image recognition. For testing, images were acquired with the use of a web camera. The accuracy of sign recognition in images is more than 80%. The real-time version of the system was implemented.
19th International Conference Computational Problems of Electrical Engineering, 2018
The article describes a system of fatigue symptoms detection of a driver, based on his behavior o... more The article describes a system of fatigue symptoms detection of a driver, based on his behavior observed with a camera. The software was written in C++. Selected functions from OpenCV and Dlib libraries were used. We analyzed the following symptoms indicating driver fatigue: blinking, yawning, turning the head, falling head forward and to the side. Experiments were performed using YawDD database. Satisfactory effectiveness of fatigue symptoms detection was achieved. The effectiveness of blink detection was 61%. For the rest of symptoms the detection accuracy was about 86%.
2020 IEEE 21st International Conference on Computational Problems of Electrical Engineering (CPEE), 2020
The article presents an algorithm for visual inspection of traffic intensity. At first, the acqui... more The article presents an algorithm for visual inspection of traffic intensity. At first, the acquisition process of video material from a road camera is described. Then the algorithm for processing and analyzing images from the recorded video material is presented. Software was prepared in MATLAB environment. Algorithm tests were conducted in real conditions, at different times of the day, different atmospheric conditions and different levels of traffic intensity. Test results show that in good working conditions the vehicle counting accuracy was 95.6%. When the sun shined in the camera lens the counting accuracy decreased to 87.2%. The smallest accuracy 83.2% was noted for traffic jams.
Experimental Setup Five users, at the age of 23, 25 31, 42, and 46 participated in the experiment... more Experimental Setup Five users, at the age of 23, 25 31, 42, and 46 participated in the experiment. Users sat comfortably in a chair. A green LED of a 1cm diameter was placed at a distance of about 1 meter from the eyes of a person. EEG signals were recorded using g.USBAmp with 16 active electrodes. Users were stimulated with flickering LED light of frequencies: 5Hz, 6Hz, 7Hz and 8Hz. The stimulation lasted 30 seconds. All sessions took place at the same time of the day to avoid circadian influences on the measurements. The electrodes were placed according to the international 10-20 system at positions: O2, AF3, AF4, P4, P3, F4, Fz, F3, FCz, Pz, C4, C3, CPz, Cz, Oz, O1. EEG sampling frequency was 256Hz. The signals were recorded using a Butterworth bandpass filter (0.1-100Hz) and notch filter (48-52Hz) to correct a technical artifact from the power network. Format of the Data For every user there are 4 files (for the stimuli of 5Hz, 6Hz, 7Hz and 8Hz each). Data are provided in Matlab format (*.mat) as X variable containing raw EEG signals: the 16 EEG potentials acquired in order: O2, AF3, AF4, P4, P3, F4, Fz, F3, FCz, Pz, C4, C3, CPz, Cz, Oz, O1.
2018 IEEE International Symposium on Medical Measurements and Applications (MeMeA), 2018
The presurgical evaluation patients for resective epilepsy surgery require localization of the ep... more The presurgical evaluation patients for resective epilepsy surgery require localization of the epileptogenic cortical zone (EZ). The detection and analysis of interictal and ictal epileptiform spikes is of major importance for identifying this area. The “irritative zone” of cortex with interictal spikes are usually revealed intraoperatively during acute electrocorticogram (ECoG). Since ECoG recordings cannot be completely visually reviewed in a reasonable amount of time, computer algorithms for automatic detection of seizures and spikes were developed. In this article we present a method of spike detection in ECoG signal using lagged phase space (LPS). Vectors in lagged phased space are treated as features. For spike detection we used expectation-maximization (EM) clustering algorithm. In this way we obtained quite high detection accuracy 96.4%.
2017 18th International Conference on Computational Problems of Electrical Engineering (CPEE), 2017
The article presents the use of genetic algorithm (GA) to select and classify ERD/ERS patterns. O... more The article presents the use of genetic algorithm (GA) to select and classify ERD/ERS patterns. One hundred twenty eight channel EEG signal was used in the experiments. The signal was recorded for 40 people, during the process of imagining right and left hand movements. Feature extraction was performed using frequency analysis (FFT) with the resolution of 1Hz. So the features were spectral lines associated with particular electrodes. The selection of features, calculated for all people, was made with GA. The fitness function used in GA was EEG signal classification error calculated using LDA classifier and 5-CV test. The average accuracy of the classification for all people in 8–30Hz band was 0.85, while for the top 10 results 0.92.
The aim of the article is to provide a systematic presentation of basic tools that are most commo... more The aim of the article is to provide a systematic presentation of basic tools that are most commonly used to analyze electroencephalography signals (EEG) in brain–computer interfaces for detection of steady-state visually evoked potentials (SSVEP). We use a database of EEG signals containing SSVEP and demonstrate the desirability of the use of selected methods, showing their benefits. Methods such as independent components analysis (ICA), frequency analysis (DFT), and time-frequency analysis (STFT) are presented. For SSVEP, the features of EEG signal should be stable with time. Short-Time Fourier Transform (STFT) allows to confirm this stability. Independent Component Analysis is used to extract pure SSVEP components. The advantages of each method are described and the obtained results are discussed. Further, source location by the use of low-resolution electromagnetic tomography algorithm is demonstrated.
Although the psychophysiological signs of fatigue are well known, automatic methods for the detec... more Although the psychophysiological signs of fatigue are well known, automatic methods for the detection of fatigue in employees in specific working conditions are still lacking. Many people do repetitive work on computers and become fatigued; therefore, the detection of fatigue in employees can help prevent accidents and increase their work efficiency. In this article, we propose an algorithm for the effective detection of fatigue which is based only on electrooculographic (EOG) signal. Three features were assessed: blink duration, blink amplitude, and time between blinks. To cause fatigue, the ${N}$ -back test, lasting for 60 minutes, was carried out. The article presents the research results for 24 users. The effectiveness of the proposed system was measured by the accuracy of classification. The average classification accuracy was 0.93 for user-dependent mode and 0.89 for user-independent mode. The results of the conducted experiments indicated that assessing the three proposed features can help in the effective detection of fatigue in users.
2019 IEEE 20th International Conference on Computational Problems of Electrical Engineering (CPEE)
The article presents a gender identification based on speech signal with supervised machine learn... more The article presents a gender identification based on speech signal with supervised machine learning implementation. At first, a database of speech signals in Polish language was collected. Next, a set of features from audio signal were calculated. The features were farther used to train a neural network. Audio signal processing and implementation of the neural network were performed in Python, and the calculation of features in the R language. Neural network training process was carried out using only CPU, then CPU with GPU and the times of the programs execution were compared. The obtained accuracy of gender recognition was 92.4%. The use of GPU accelerated the network learning process several times.
This article describes a vision system that uses deep learning to recognize 24 static signs of th... more This article describes a vision system that uses deep learning to recognize 24 static signs of the American Sign Alphabet in real time. As part of the project, images of signs from four publicly available databases were used as a training set. A DenseNet was implemented for image recognition. For testing, images were acquired with the use of a web camera. The accuracy of sign recognition in images is more than 80%. The real-time version of the system was implemented.
19th International Conference Computational Problems of Electrical Engineering, 2018
The article describes a system of fatigue symptoms detection of a driver, based on his behavior o... more The article describes a system of fatigue symptoms detection of a driver, based on his behavior observed with a camera. The software was written in C++. Selected functions from OpenCV and Dlib libraries were used. We analyzed the following symptoms indicating driver fatigue: blinking, yawning, turning the head, falling head forward and to the side. Experiments were performed using YawDD database. Satisfactory effectiveness of fatigue symptoms detection was achieved. The effectiveness of blink detection was 61%. For the rest of symptoms the detection accuracy was about 86%.
2020 IEEE 21st International Conference on Computational Problems of Electrical Engineering (CPEE), 2020
The article presents an algorithm for visual inspection of traffic intensity. At first, the acqui... more The article presents an algorithm for visual inspection of traffic intensity. At first, the acquisition process of video material from a road camera is described. Then the algorithm for processing and analyzing images from the recorded video material is presented. Software was prepared in MATLAB environment. Algorithm tests were conducted in real conditions, at different times of the day, different atmospheric conditions and different levels of traffic intensity. Test results show that in good working conditions the vehicle counting accuracy was 95.6%. When the sun shined in the camera lens the counting accuracy decreased to 87.2%. The smallest accuracy 83.2% was noted for traffic jams.
Experimental Setup Five users, at the age of 23, 25 31, 42, and 46 participated in the experiment... more Experimental Setup Five users, at the age of 23, 25 31, 42, and 46 participated in the experiment. Users sat comfortably in a chair. A green LED of a 1cm diameter was placed at a distance of about 1 meter from the eyes of a person. EEG signals were recorded using g.USBAmp with 16 active electrodes. Users were stimulated with flickering LED light of frequencies: 5Hz, 6Hz, 7Hz and 8Hz. The stimulation lasted 30 seconds. All sessions took place at the same time of the day to avoid circadian influences on the measurements. The electrodes were placed according to the international 10-20 system at positions: O2, AF3, AF4, P4, P3, F4, Fz, F3, FCz, Pz, C4, C3, CPz, Cz, Oz, O1. EEG sampling frequency was 256Hz. The signals were recorded using a Butterworth bandpass filter (0.1-100Hz) and notch filter (48-52Hz) to correct a technical artifact from the power network. Format of the Data For every user there are 4 files (for the stimuli of 5Hz, 6Hz, 7Hz and 8Hz each). Data are provided in Matlab format (*.mat) as X variable containing raw EEG signals: the 16 EEG potentials acquired in order: O2, AF3, AF4, P4, P3, F4, Fz, F3, FCz, Pz, C4, C3, CPz, Cz, Oz, O1.
2018 IEEE International Symposium on Medical Measurements and Applications (MeMeA), 2018
The presurgical evaluation patients for resective epilepsy surgery require localization of the ep... more The presurgical evaluation patients for resective epilepsy surgery require localization of the epileptogenic cortical zone (EZ). The detection and analysis of interictal and ictal epileptiform spikes is of major importance for identifying this area. The “irritative zone” of cortex with interictal spikes are usually revealed intraoperatively during acute electrocorticogram (ECoG). Since ECoG recordings cannot be completely visually reviewed in a reasonable amount of time, computer algorithms for automatic detection of seizures and spikes were developed. In this article we present a method of spike detection in ECoG signal using lagged phase space (LPS). Vectors in lagged phased space are treated as features. For spike detection we used expectation-maximization (EM) clustering algorithm. In this way we obtained quite high detection accuracy 96.4%.
2017 18th International Conference on Computational Problems of Electrical Engineering (CPEE), 2017
The article presents the use of genetic algorithm (GA) to select and classify ERD/ERS patterns. O... more The article presents the use of genetic algorithm (GA) to select and classify ERD/ERS patterns. One hundred twenty eight channel EEG signal was used in the experiments. The signal was recorded for 40 people, during the process of imagining right and left hand movements. Feature extraction was performed using frequency analysis (FFT) with the resolution of 1Hz. So the features were spectral lines associated with particular electrodes. The selection of features, calculated for all people, was made with GA. The fitness function used in GA was EEG signal classification error calculated using LDA classifier and 5-CV test. The average accuracy of the classification for all people in 8–30Hz band was 0.85, while for the top 10 results 0.92.
The aim of the article is to provide a systematic presentation of basic tools that are most commo... more The aim of the article is to provide a systematic presentation of basic tools that are most commonly used to analyze electroencephalography signals (EEG) in brain–computer interfaces for detection of steady-state visually evoked potentials (SSVEP). We use a database of EEG signals containing SSVEP and demonstrate the desirability of the use of selected methods, showing their benefits. Methods such as independent components analysis (ICA), frequency analysis (DFT), and time-frequency analysis (STFT) are presented. For SSVEP, the features of EEG signal should be stable with time. Short-Time Fourier Transform (STFT) allows to confirm this stability. Independent Component Analysis is used to extract pure SSVEP components. The advantages of each method are described and the obtained results are discussed. Further, source location by the use of low-resolution electromagnetic tomography algorithm is demonstrated.
Uploads
Papers by Andrzej Majkowski